Twitter Geo-Sentiment Analysis
During Corona Crisis

The Idea

Sentiment analysis of social media data is a powerful tool. I came up with this idea to analyze twitter data using corona keywords. This analysis allows to create temporal and spatial snapshots of the public opinion regarding topics related to the Corona-Crisis such as social distancing, curfews or healthcare occupancy. To limit the number of languages for the sentiment analysis, I first started out to analyze tweets captured in Germany.

Workflow

The Data

Wrap-up

Sentiment analysis of tweets related to the Corona-Crisis revealed some interesting insights of public's opinion. We can visualize differences in the public's sentiment, the tweet locations and their quantity over time. Fluctuations in both polarity and subjectivity of the tweet are observed over time. The mean polarity seems to be slightly positive which might be surprising considering the nature of social media. Interestingly, on 24.05. there is a slight trend towards a negative polarity visible. However, the sensitivity of this type of sentiment analysis is limited and probably requires more advanced natural language processing approaches.

About

Profile Picture

Mehrshad Pakdel

Hey there! After my PhD in biochemistry and first professional experience as a Consultant for Data Analytics, I am thrilled about data science and coding. As an aspiring data scientist, I am excited to learn new Python techniques and frameworks. If you have questions about this project, please don't hesitate to contact me via LinkedIn.


Legal Notice